Imagen de Portada

TensorHub: Scalable and Elastic Weight Transfer for LLM RL Training

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Ye, Chenhao, Zhang, Huaizheng, Han, Mingcong, Zhong, Baoquan, Li, Xiang, Chen, Qixiang, Zhang, Xinyi, Zhang, Weidong, Jiang, Kaihua, Zhang, Wang, Sun, He, Xiao, Wencong, Arpaci-Dusseau, Andrea C., Arpaci-Dusseau, Remzi H.
Formato:	Preprint
Publicado:	2026
Materias:	Distributed, Parallel, and Cluster Computing Artificial Intelligence
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

Bodega: Serving Linearizable Reads Locally from Anywhere at Anytime via Roster Leases
por: Hu, Guanzhou, et al.
Publicado: (2025)

A Unified, Practical, and Understandable Model of Non-transactional Consistency Levels in Distributed Replication
por: Hu, Guanzhou, et al.
Publicado: (2024)

Crossword: Adaptive Consensus for Dynamic Data-Heavy Workloads
por: Hu, Guanzhou, et al.
Publicado: (2025)

Foreactor: Exploiting Storage I/O Parallelism with Explicit Speculation
por: Hu, Guanzhou, et al.
Publicado: (2024)

Getting the MOST out of your Storage Hierarchy with Mirror-Optimized Storage Tiering
por: Tu, Kaiwei, et al.
Publicado: (2025)

Zerrow: True Zero-Copy Arrow Pipelines in Bauplan
por: Dai, Yifan, et al.
Publicado: (2025)

Don't Let AI Agents YOLO Your Files: Shifting Information and Control to Filesystems for Agent Safety and Autonomy
por: Zhong, Shawn Wanxiang, et al.
Publicado: (2026)

Tidying Up the Address Space
por: Banakar, Vinay, et al.
Publicado: (2025)

OBASE: Object-Based Address-Space Engineering to Improve Memory Tiering
por: Banakar, Vinay, et al.
Publicado: (2026)

StreamRL: Scalable, Heterogeneous, and Elastic RL for LLMs with Disaggregated Stream Generation
por: Zhong, Yinmin, et al.
Publicado: (2025)

Efficient Pre-Training of LLMs via Topology-Aware Communication Alignment on More Than 9600 GPUs
por: He, Guoliang, et al.
Publicado: (2025)

DistFlow: A Fully Distributed RL Framework for Scalable and Efficient LLM Post-Training
por: Wang, Zhixin, et al.
Publicado: (2025)

ElasWave: An Elastic-Native System for Scalable Hybrid-Parallel Training
por: Kang, Xueze, et al.
Publicado: (2025)

Accurate and Fast Approximate Graph Pattern Mining at Scale
por: Arpaci-Dusseau, Anna, et al.
Publicado: (2024)

Federated Learning Using Coupled Tensor Train Decomposition
por: Zhang, Xiangtao, et al.
Publicado: (2024)

Role-Based Fault Tolerance System for LLM RL Post-Training
por: Chen, Zhenqian, et al.
Publicado: (2025)

Fast LLM Post-training via Decoupled and Fastest-of-N Speculation
por: Cheng, Rongxin, et al.
Publicado: (2025)

AnchorTP: Resilient LLM Inference with State-Preserving Elastic Tensor Parallelism
por: Xu, Wendong, et al.
Publicado: (2025)

Spatial-Temporal Federated Learning for Lifelong Person Re-identification on Distributed Edges
por: Zhang, Lei, et al.
Publicado: (2022)

Laminar: A Scalable Asynchronous RL Post-Training Framework
por: Sheng, Guangming, et al.
Publicado: (2025)

DynaTrain: Fast Online Parallelism Switching for Elastic LLM Training
por: Wang, Yuanqing, et al.
Publicado: (2026)

Rubick: Exploiting Job Reconfigurability for Deep Learning Cluster Scheduling
por: Zhang, Xinyi, et al.
Publicado: (2024)

LiveR: Fine-Grained Elasticity via Live Reconfiguration for Model Training
por: Liu, Haoyuan, et al.
Publicado: (2026)

Schedule-Level Shared-Prefix Reuse for LLM RL Training
por: Li, Pengbo, et al.
Publicado: (2026)

ROSE: Rollout On Serving GPUs via Cooperative Elasticity for Agentic RL
por: Gao, Wei, et al.
Publicado: (2026)

TACO: Efficient Communication Compression of Intermediate Tensors for Scalable Tensor-Parallel LLM Training
por: Liu, Man, et al.
Publicado: (2026)

Cabinet: Dynamically Weighted Consensus Made Fast
por: Zhang, Gengrui, et al.
Publicado: (2025)

ElasticMoE: An Efficient Auto Scaling Method for Mixture-of-Experts Models
por: Singh, Gursimran, et al.
Publicado: (2025)

Elastic Data Transfer Optimization with Hybrid Reinforcement Learning
por: Swargo, Rasman Mubtasim, et al.
Publicado: (2025)

NeutronTP: Load-Balanced Distributed Full-Graph GNN Training with Tensor Parallelism
por: Ai, Xin, et al.
Publicado: (2024)

HetRL: Efficient Reinforcement Learning for LLMs in Heterogeneous Environments
por: He, Yongjun, et al.
Publicado: (2025)

RL in the Wild: Characterizing RLVR Training in LLM Deployment
por: Zhou, Jiecheng, et al.
Publicado: (2025)

Polar: Agentic RL on Any Harness at Scale
por: Xu, Binfeng, et al.
Publicado: (2026)

Synergistic Tensor and Pipeline Parallelism
por: Qi, Mengshi, et al.
Publicado: (2025)

WOC: Dual-Path Weighted Object Consensus Made Efficient
por: Fonseca, Tanisha, et al.
Publicado: (2025)

PromptTuner: SLO-Aware Elastic System for LLM Prompt Tuning
por: Gao, Wei, et al.
Publicado: (2026)

SparseRL-Sync: Lossless Weight Synchronization with ~100x Less Communication
por: Hu, Lucas, et al.
Publicado: (2026)

Lazarus: Resilient and Elastic Training of Mixture-of-Experts Models
por: Wu, Yongji, et al.
Publicado: (2024)

Runtime-Orchestrated Second-Order Optimization for Scalable LLM Training
por: Lu, Yishun, et al.
Publicado: (2026)

FlowMoE: A Scalable Pipeline Scheduling Framework for Distributed Mixture-of-Experts Training
por: Gao, Yunqi, et al.
Publicado: (2025)